NVIDIA has just released Nemotron 3.5 Bolt E4B 4GB, an open small language model for fast conversational agents. It targets voice assistants, game NPCs and robot interfaces running locally on RTX, Jetson and Spark hardware. The model is post-trained from Google’s Gemma 4 E4B using reinforcement learning only. For developers, this is agent behavior in…
Moonshot AI has just released Kimi K3.1 Pro, the first model in its K3.1 series. Kimi K3.1 Pro is a 3T-class open-weight* model with a 1-million-token context window. Moonshot team positions it against the strongest proprietary models on coding, agentic and knowledge benchmarks. For developers, it is cheaper API pricing…
Building a startup means making decisions before you have all the answers. When should you scale? How do you know you really have product-market fit? What does it take to land enterprise customers? How do you build the right team? And when you finally get in front of an investor, how do you make sure…
Kyutai has released CADRE, open causal diffusion representation autoencoders for interactive world models. The Kyutai CADRE video autoencoder targets one problem: action lag. Many world models borrow autoencoders from text-to-video generators. Those autoencoders pack many frames into one latent, so every action waits. Per the research paper, CADRE world models run at 20 fps on…
TechCrunch Founder Summit 2026 is coming to SoWa Power Station in Boston on November 4, bringing founders, investors, and operators together for a full day of candid conversations, practical sessions, live pitching, networking, and smaller roundtables built for deeper discussion. The roundtables are where attendees can get especially specific. Bring your questions, share what you…
On Friday, OpenAI said that it will train its models using cached copies and simulated versions of the internet instead of the live web because the models are too unpredictable to be trusted with live internet access. The frontier lab announced its new policy two weeks after its rival, Anthropic, began limiting its newest and,…
Getting initial customers as a startup can be one of the most grueling challenges in business. Even when companies manage to snag some early interest, there’s no guarantee those customers will stick around. At TechCrunch Disrupt last week, Elia Wallen, founder and CEO of travel startup Engine, Gamma founder and CEO Grant Lee, and Crystal…
Earlier this week, a video of Nvidia CEO Jensen Huang playing with a flock of waddling Microducks during his keynote at the company’s GTC Berlin conference went viral. The robots, stomping their little feet and making tiny R2-D2 noises, sat, fell down dead for a “nap,” and summersaulted on voice command. They were so darn…
Product-market fit — when a startup’s product sells well — has become even harder to find in this AI era. Despite startups achieving record-fast growth, not all customers are sticky and products can be replicated more easily than ever. Anu Bharadwaj, a partner at Iconiq; Rajeev Dham, a partner at Sapphire Ventures; and Rahul Vohra,…
Product-market fit — when a startup’s product sells well — has become even harder to find in this AI era. Despite startups achieving record-fast growth, not all customers are sticky, and products can be replicated more easily than ever. Anu Bharadwaj, a partner at Iconiq; Rajeev Dham, a partner at Sapphire Ventures; and Rahul Vohra,…
